Dose-dependent regulation of microbial activity on
sinking particles by polyunsaturated aldehydes:
Implications for the carbon cycle
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Edwards, Bethanie R., Kay D. Bidle, and Benjamin A. S. Van
Mooy. “Dose-Dependent Regulation of Microbial Activity on
Sinking Particles by Polyunsaturated Aldehydes: Implications for
the Carbon Cycle.” Proc Natl Acad Sci USA 112, no. 19 (April 27,
2015): 5909–5914.
As Published
National Academy of Sciences (U.S.)
Final published version
Thu May 26 18:50:24 EDT 2016
Citable Link
Terms of Use
Article is made available in accordance with the publisher's policy
and may be subject to US copyright law. Please refer to the
publisher's site for terms of use.
Detailed Terms
Dose-dependent regulation of microbial activity on
sinking particles by polyunsaturated aldehydes:
Implications for the carbon cycle
Bethanie R. Edwardsa,b, Kay D. Bidlec, and Benjamin A. S. Van Mooya,1
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543; bDepartment of Earth, Atmospheric,
and Planetary Science, Massachusetts Institute of Technology, Cambridge, MA 02139; and cDepartment of Marine and Coastal Sciences, Rutgers University,
New Brunswick, NJ 08901
polyunsaturated aldehydes sinking particles particle-associated
bacteria marine carbon cycle bioactivity hypothesis
interest. The stage for discovery of these molecules was set in the
1990s, when a group of researchers advanced the “paradox of
diatom-copepod interactions”—the observation that copepods,
which prey on diatoms, exhibited decreased reproductive success
when exclusively fed diatoms (6–8). Miralto et al. (9) later purified
PUAs from diatom cultures and observed arrested embryogenesis
of copepod eggs that were exposed to these compounds. PUA
production is now a well-characterized stress surveillance response
to wounding during grazing and to nutrient depletion, both of
which are bloom termination mechanisms (10, 11).
It has been proposed that PUAs also mediate phytoplankton
bloom dynamics by impacting other members of the marine
planktonic community, aside from zooplankton. In culture conditions, many eukaryotic phytoplankton experience a decrease in
growth rate when exposed to PUAs (12, 13). Isolated bacterial
strains demonstrate a varied response to PUAs, whereby diatomassociated isolates are generally unaffected, whereas other strains
exhibit either dose-dependent decreases or increases in growth
rate in response to PUAs (14). Although many culture studies
have been conducted on zooplankton, phytoplankton, and bacterial isolates—all important players in the microbial loop—
there have been few attempts to study the impact of PUAs on
these trophic levels in natural marine ecosystems under in situ
condition (15, 16). Surveys of water column concentrations of
PUAs suggest that concentrations are generally much lower than
levels required to elicit responses in phytoplankton, zooplankton,
lanktonic microbes in the world’s oceans play a major role in
the global carbon cycle. Through photosynthesis, phytoplankton convert carbon dioxide into particulate organic carbon
(POC), which then has the potential to sink to the deep sea. This
process is opposed by zooplankton and heterotrophic bacteria,
which, as agents of respiration, degrade organic matter and convert
it back into carbon dioxide. In addition to respiring POC, heterotrophic bacteria that are associated with sinking POC use membrane-bound ectohydrolytic enzymes that affect the disaggregation
of POC into smaller nonsinking particles and dissolved organic
carbon (DOC) (1). Despite the long-recognized role of particleassociated bacteria (2), the relationships between the activities of
these bacteria, particle properties, and the time and depth scales of
sinking POC degradation, disaggregation, and respiration remain
poorly constrained. Although the molecular-level composition of
sinking POC has been used to identify its phytoplanktonic sources
and to constrain the timescales of its degradation by heterotrophic
bacteria (3–5), none of these studies has accounted for the potential
impacts of bioactive molecules within sinking POC on the activities
of particle-associated bacteria.
Diatoms are key members of the phytoplanktonic communities across the world’s ocean and are known to produce a large
diversity of organic molecules, including many that are bioactive.
Polyunsaturated aldehydes (PUAs) have received particular
Phytoplankton live in the sunlit surface waters of the ocean,
and through photosynthesis they convert atmospherically derived carbon dioxide into their biomass. A fraction of this
biomass sinks into the darker depths where it is colonized by
bacteria that turn it back into carbon dioxide through respiration. Thus, phytoplankton–bacteria interactions effectively
transport carbon dioxide from the atmosphere deep into the
ocean. We discovered that the biomass of some phytoplankton
contains bioactive molecules that stimulate these associated
bacteria, resulting in respiration of phytoplankton biomass at
shallower depths. Given that the ocean mixes gradually over
time, carbon dioxide released by bacteria at shallower depths
returns to the surface more quickly and thereby is “sequestered” from the atmosphere for a shorter duration.
Author contributions: B.R.E., K.D.B., and B.A.S.V.M. designed research; B.R.E. and B.A.S.V.M.
performed research; B.R.E. and B.A.S.V.M. analyzed data; and B.R.E., K.D.B., and B.A.S.V.M.
wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
To whom correspondence should be addressed. Email: [email protected]
This article contains supporting information online at
PNAS | May 12, 2015 | vol. 112 | no. 19 | 5909–5914
Diatoms and other phytoplankton play a crucial role in the global
carbon cycle, fixing CO2 into organic carbon, which may then be
exported to depth via sinking particles. The molecular diversity of this
organic carbon is vast and many highly bioactive molecules have
been identified. Polyunsaturated aldehydes (PUAs) are bioactive on
various levels of the marine food web, and yet the potential for these
molecules to affect the fate of organic carbon produced by diatoms
remains an open question. In this study, the effects of PUAs on the
natural microbial assemblages associated with sinking particles were
investigated. Sinking particles were collected from 150 m in the water
column and exposed to varying concentrations of PUAs in dark incubations over 24 h. PUA doses ranging from 1 to 10 μM stimulated
respiration, organic matter hydrolysis, and cell growth by bacteria
associated with sinking particles. PUA dosages near 100 μM appeared
to be toxic, resulting in decreased bacterial cell abundance and metabolism, as well as pronounced shifts in bacterial community composition. Sinking particles were hot spots for PUA production that
contained concentrations within the stimulatory micromolar range in
contrast to previously reported picomolar concentrations of these
compounds in bulk seawater. This suggests PUAs produced in situ
stimulate the remineralization of phytoplankton-derived sinking organic matter, decreasing carbon export efficiency, and shoaling the
average depths of nutrient regeneration. Our results are consistent
with a “bioactivity hypothesis” for explaining variations in carbon
export efficiency in the oceans.
Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved March 24, 2015 (received for review December 1, 2014)
Exposure to PUAs Affects Changes in the Rates of Sinking POC
Remineralization. We tested the linkages between PUAs and the
rate of organic matter remineralization by particle-associated bacteria at six stations across the North Atlantic Ocean (Fig. S1 and
Table S1): three stations in the Sargasso Sea (SS), two in the
temperate western North Atlantic (TWNA), and one in the Subarctic North Atlantic (SANA). Our experimental methods centered
on collecting sinking particles, incubating particles in the presence
of exogenous PUAs (mixture of heptadienal, octadienal, and decadienal) at a range of concentrations, and assessing changes in organic matter respiration, hydrolytic enzyme activity, bacterial cell
abundance, bacterial production rates, and bacterial community
structure. The absolute values of these parameters varied considerably between stations (Fig. S2). To remove between-stations
variability, we divided the average values of the PUA-amended
treatments by the average of the no-amendment controls from each
corresponding station. These control-normalized data showed
strikingly similar responses by particle-associated bacteria to PUA
treatments across this ocean basin (Figs. 1 and 2). We then asked
whether there were differences between the controls and the incubations amended with different concentrations of PUAs using a
series of Wilcoxon ranked-sum statistical tests.
In general, the addition of exogenous PUAs at lower concentrations led to stimulated rates of bacterial organic matter
remineralization. The average respiration rates in the 1 and 10 μM
treatments were approximately double that observed in the control treatment (Fig. 1). The average respiration rate in the 100 μM
treatment did not differ from the control. Enzyme activity assays
revealed enhanced alkaline phosphatase (APase) and lipase activity compared with the control over the same stimulatory range
of concentrations observed for respiration rates (Fig. 2). Average
APase activity in the 1 and 10 μM treatments was quadruple that
of the controls. The average lipase activity was one-and-a-half
Values Normalized to the Control
Bact. Cell Count
Bact. Production
1 μM
10 μM
100 μM
Fig. 1. Average effects of PUA treatments on respiration (n = 18), production (n = 9), and abundance (n = 17) of particle-associated cells, presented as the normalized ratio to the no-amendment control incubations.
Data are derived from triplicate incubations at six stations. To isolate and
compare the effects of the amendments across stations, results from the
incubations at each station were normalized by dividing by the average
value of the no-amendment control. With the between-station variability
removed, differences between treatments and the control were identified
using Wilcoxon statistical tests. The red values above the data report the
average values for each treatment across all experiments and denote statistical difference from the control, P < 0.05 (Wilcoxon rank sum).
5910 |
Enzyme Acitivy Normalized to the Control
or bacteria (17, 18). Consequently, the impact of PUAs on the
marine carbon cycle remains an open question.
1 μM
10 μM
100 μM
Fig. 2. Average effects of PUA amendments on APase, lipase, aminopeptidase, and α-glucosidase activity (n = 18 for all), presented as the normalized
ratio to the no-amendment control incubations. Statistical analyses conducted as described for Fig. 1. The red values above the data report the
average values for each treatment across all experiments and denote statistical difference from the control, P < 0.05 (Wilcoxon rank sum).
times that of the control in the 1 and 10 μM treatments. Lipase
activity was significantly lower than the control in the 100 μM
treatments. Peptidase activity was significantly lower than the
control in the 10 and 100 μM treatments. However, α-glucosidase
activity did not significantly deviate from the control in any of
the treatments.
The changes in the rates of organic matter remineralization in
response to PUAs were reflected in the growth of particle-associated bacteria. Bacterial cell abundances in the 10 μM treatments
were ∼50% greater than in the controls (Fig. 1), whereas bacterial
production was about 20% higher (Fig. 1). Similar responses in
these signals, although of lesser magnitude, were observed in the
1 μM PUA treatments, hinting at a dose-dependent growth response to PUAs. In contrast to the 1 and 10 μM treatments,
bacterial cell abundance was significantly lower than the control in
the 100 μM PUA treatments, pointing to an inhibitory threshold
between 10 and 100 μM. The average bacterial production rates
were also generally lower in the 100 μM treatment compared with
the control, but this effect was not statistically significant because
of geographic variability; the two TWNA sites showed almost
complete inhibition of bacterial production, whereas SS3 showed
stimulation (Fig. S2).
Sinking Particles are Hot Spots for PUA Production. Sinking particles
were also collected for PUA analysis by high-performance liquid
chromatography–UV–multistage mass spectrometry (HPLC-UVMSn). Decadienal was clearly observed in the sinking particles at
TWNA2 before incubation (i.e., t = initial; Fig. S3); based on POC
content of these sinking particles and a previously published relationship between POC and volume of diatom-derived marine snow
particles (19), the in situ concentration of decadienal within sinking
particles was estimated to be 26 μM (Fig. 3 and Table S2). This
concentration was comparable to the stimulatory range of concentrations observed in our incubation experiments. Additionally, significant production of PUAs of multiple chain lengths was observed
in the no-amendment controls after 24 h of incubation (i.e., t = final)
in all three sampling regions (Fig. 3 and Table S2); heptadienal,
octadienal, and decadienal reached concentrations as high as 10.4,
12.9, and 34.0 μM, respectively.
Shift in Bacterial Community upon Exposure to PUAs. Changes in the
community structure of particle-associated bacteria were assessed
by using automated ribosomal intergenic spacer analysis (ARISA)
Edwards et al.
N.D. N.D.
TWNA2 t=0
N.D. N.D.
Fig. 3. Box plots displaying the average concentration, range of concentrations, and SD from the mean of heptadienal (gray), octadienal (blue), and
decadienal (green) [in micromoles per liter particle] for the t = final control
treatments from incubation experiments at SS1, TWNA1, TWNA2, and
SANA1 and the t = initial from TWNA2 (n = 3 for all). These values were
estimated from the PUA concentration within the incubations, the POC
content of the trap material, and the relationship between POC and volume
of diatom-derived marine sinking particulate matter published by Brzezinski
et al. (19). N.D., not determined.
at one station in the TWNA (TWNA1) and at one station in the
Sargasso Sea (SS3). The presence/absence of each operational
taxonomic unit (OTU) identified in the various treatments was
analyzed with multidimensional scaling to describe the variations in
community structure. The particle-associated bacteria communities
from the control, 1 μM, and 10 μM treatments formed two distinct
clusters based on geographic location (Fig. 4), indicative of negligible impacts on bacterial community structure at these low PUA
concentrations. In contrast, the 100 μM treatments led to dramatic
changes in community structure vs. the other treatments.
An ARISA clone library database was used to assign a putative
identity to each OTU, and we calculated the relative abundance
of the following bacterial phyla/classes: Actinobacteria, Bacteroidetes, Cyanobacteria, Deferribacteres, Firmicutes, α-Proteobacteria, β-Proteobacteria, δ-Proteobacteria, Ɛ-Proteobacteria,
and γ-Proteobacteria (Fig. S4). The effects of PUAs on individual
clades were assessed using Spearman’s rank correlations. Abundances of most groups did not show a significant correlation with
the amount of PUA added, which includes noted particle specialists (e.g., Bacteriodetes and Firmicutes) (20, 21). At the same
time, γ-Proteobacteria were significantly negatively correlated
with PUA concentrations and Actinobacteria were significantly
positively correlated (Table S3).
Doses of PUAs ranging from 1 to 10 μM stimulated organic
matter respiration on sinking particles (Fig. 1). Enhanced organic
matter respiration in these treatments supported bacterial growth,
as suggested by parallel increases in bacterial cell abundance and
bacterial production (Fig. 1). Stimulatory concentrations of 1–
10 μM agree with data reported for two cultured bacterial strains,
Eudora adriatica and Alteromonas hispanica, which showed enhanced growth rates when exposed to PUA concentrations as low
as 13 μM (14). Ribalet et al. (14) conducted incubations that
suggested that the bioactivity of PUAs is derived specifically from
their combination of carbonyl group and double bonds, precluding
the potential for PUAs to be used as a food source.
If PUAs were a consistent food source, then a relationship
between the PUA consumption ([PUA]t=initial – [PUA]t=final) and
oxygen consumption by respiration ([O2]t=initial – [O2]t=final)
would be expected. This comparison was made for the 1–10 μM
treatments in experiments conducted at TWNA1, TWNA2, and
Edwards et al.
SS3 where [PUA]t=initial and [PUA]t=final data were available
(Table S4). In three of the six comparisons, the consumption of
oxygen was greater than the drawdown in PUAs, yielding strong
evidence in those instances that the stimulation of respiration
was not driven solely by respiration of the PUA amendments
themselves. Indeed, in the 10 μM experiments at SS3, and in
most of the control experiments across the study (Fig. 3), net
PUA production was observed, which further suggests that PUAs
are not a readily accessible food source. The remaining two incubations where oxygen consumption was less than PUA drawdown were from TWNA1; this does not necessarily contradict
the results from the other four incubations because PUAs could
have been partially degraded, which would remove them from
our analytical window without incurring stoichiometric oxygen
consumption. The inconsistent relationship between respiration
and PUA consumption bolsters our interpretation that the
stimulatory effect of PUAs was not simply the result of direct
respiration of these molecules.
Our data and subsequent calculations suggest that PUA concentrations in environmental samples of sinking particles were in
the low micromolar range and comparable to the stimulatory range
in the incubation experiments (Fig. 3 and Table S2). The concentrations of PUAs within sinking particles were calculated using a
previously published POC–volume relationship for diatom-derived
marine snow particles (19), and thus there are considerable uncertainties in these concentrations. However, dissolved concentrations of PUAs from phytoplankton in North Atlantic seawater were
recently determined to be generally less than 1 pM (18). Because
the concentrations we observed in sinking particles are orders of
magnitude higher, we propose that sinking particles are hot spots
for PUA production. This idea is supported both by the direct
observation of decadienal in native (i.e., t = initial) particles from
TWNA, as well as the accumulation of PUAs to micromolar concentrations within particles from four of the six no-amendment
control treatments (t = final; Fig. 3).
The accumulation of PUAs in the incubations also suggests
that PUAs continue to be produced as particles descend into the
mesopelagic. It will be important to quantify PUAs on particles at
different depths (vs. only 150 m in this study) because the accumulation of PUAs to concentrations above ∼10 μM during transit
could potentially result in a transition between stimulatory and
Dimension 1
Fig. 4. Multidimensional scaling analysis of community structures determined using presence/absence of OTUs identified by ARISA. Data points
within blue and yellow circles are from experiments TWNA1 and SS3, respectively. Solid black circles are control treatments (n = 6). Open circles are
1 μM PUA treatments (n = 5). Solid black triangles are 10 μM treatments (n = 4).
Open triangles are 100 μM treatments (n = 6).
PNAS | May 12, 2015 | vol. 112 | no. 19 | 5911
[ PUA ] μmol Lparticle
Dimension 2
toxic effects on bacteria associated with the particles at particular
depth horizons (Fig. 1). For example, a recent large-scale study of
a diatom bloom in the North Atlantic concluded that export efficiency was low at depths above 100 m (22), but then increased
substantially at depths below 100 m (23); these results are
consistent with our data suggesting that PUA dynamics could
have complex and potentially contrasting effects on the export
depths and the biogeochemical fate of POC sinking through the
water column.
Export of POC to depth via sinking particles represents a
globally significant carbon sink. The observed stimulatory concentrations of PUAs within sinking particles could affect biogeochemical cycling by decreasing POC export efficiency (Fig. 5).
Direct PUA-enhanced respiration of sinking POC would accelerate the transfer of carbon from the organic carbon pool to the
dissolved inorganic carbon pool; to a first approximation, the
shallower this occurs, the shorter the timescales of carbon sequestration from the atmosphere (24). By increasing the hydrolysis
of POC, stimulatory concentrations of PUAs could also cause
additional transfer of carbon from the POC pool into the DOC
pool through disaggregation or dissolution. Disaggregation leads
to decreased sinking speeds and dissolution to greater rates of
microbial utilization (25, 26). All of these PUA-induced changes
cumulatively lead to shallower remineralization depths causing the
release of CO2 in waters that are more likely to be mixed to the
surface and reequilibrate with the atmosphere on shorter timescales, thus attenuating carbon sequestration in the deep sea.
Fig. 5. Conceptualization of the impact of polyunsaturated aldehydes (PUAs)
on seasonal organic carbon export from diatom blooms and subsequent
remineralization of organic matter by particle-associated bacteria. Stimulatory
concentrations of PUAs lead to enhanced regeneration of nutrients above the
winter mixing depth.
5912 |
The shoaling of remineralization depths by PUAs would also
affect greater release of inorganic nutrients from sinking particles in shallower waters (Fig. 5). The depth of mixing is 100–
450 m in the North Atlantic during the winter (27), and thus
nutrients released from sinking particles above this depth during
spring diatom blooms have the potential to fuel primary productivity in the subsequent spring. APase activity quadrupled in
our incubation experiments suggesting PUAs led to enhanced
release rates of dissolved inorganic phosphorus (Fig. 2). Enhanced lipase activity directed toward phospholipids may also
affect the liberation of dissolved phosphorus and other nutrients
(28). Phosphorus can be a limiting nutrient in regions of the
North Atlantic (29). The C:P ratios of exported particles in these
regions can be high, which could also point toward enhanced
release of phosphorus (30, 31). Furthermore, physiological
studies with diatoms in cultures and in mesocosm experiments
have shown that PUA production increases under phosphorusdepleted conditions (32, 33). Thus, there appear to be feedbacks
between PUA production and organic phosphorus remineralization, potentially leading to the release of more phosphorus
above the winter mixed layer, and thereby affecting greater rates
of primary production on interannual and basinwide scales.
Although PUAs may stimulate the recycling of phosphorous,
the decreases in peptidase activity (Fig. 2) might indicate a decrease
in biogenic Si remineralization on particles because diatom frustules are covered by glycoproteins that biochemically protect frustules from contact with seawater that is undersaturated in dissolved
silica. Thus, decreased peptidase activity directed toward these
glycoproteins could result in lower dissolution of biogenic silica
(BSi) in the euphotic zone (34, 35), a process that supports ∼60%
of global BSi production (36). Preferential recycling of phosphorous
over silica could play a role in phytoplankton community succession; in the North Atlantic, diatom blooms are followed by blooms
of coccolithophores, which do not have a silica requirement (37). It
should be noted that the link between peptidase activity and BSi
dissolution does not appear to be universal across all diatomassociated strains of bacteria (38). Thus, the impact of PUAs on BSi
regeneration and the potential link with enhanced organic phosphorus cycling remain to be fully elucidated.
The variable response of different enzyme activities to different PUA concentrations might belie a connection between
PUAs and the biochemical composition of sinking particles.
Although the ectoenzymatic hydrolysis measurements were
based on a few well-established model substrates routinely used
in microbial ecology since the 1980s (39), they likely do not
encapsulate the complete enzymatic response to the molecularly diverse organic matter in sinking particles. Future studies
could focus on how the addition of PUAs to sinking particles
alters the biochemical composition of POC and DOC, as well as
the exchange of classes of biochemicals between the various
particulate and dissolved pools. However, our current data
clearly show enhanced rates of respiration when particles were
amended with ecologically relevant concentrations of PUAs.
Notably, respiration represents the ultimate utilization and
complete remineralization of organic matter to CO2 regardless
of the biochemical composition of the particles or the enzyme
activities of particle-associated bacteria.
Our finding that the addition of high doses of PUAs significantly altered the community structure of particle-associated bacteria
suggests that PUAs could play a role in bacterial community succession on sinking particles as PUA levels accumulate. Indeed,
100 μM PUA treatments exhibited dramatic shifts in community
structure combined with significant decreases in bacterial cell
abundance (Figs. 1 and 4). In contrast, lower levels of PUA
amendments in the 1–10 μM range stimulated bacterial metabolic
activity (cell abundance, production, and APase activity) (Figs. 1
and 2), while affecting much more subtle shifts in resident bacterial
community, which included phyla and classes known to contain
Edwards et al.
This is one of the first reports showing that a specific class of
bioactive molecules from phytoplankton impacts the activity and
community structure of natural bacterial communities associated
with sinking particles. We observed consistent dose-dependent
bioactivity of PUAs in six iterations of the same incubation experiment across three different regions of the North Atlantic.
Higher respiration rates, APase activity, lipase activity, and bacterial growth were observed over a stimulatory range of PUA exposure (1–10 μM) and for generally similar bacterial communities.
PUAs at higher concentrations tended to have inhibitory effects
and induced dramatic shifts in the bacterial community structure,
demonstrating that PUAs may play a role in bacterial community succession on sinking particles. Decadienal concentrations
Edwards et al.
comparable to the observed stimulatory range were observed
within sinking particles collected in the TWNA and PUAs accumulated in incubations at other locations, suggesting that
sinking particles are hot spots for PUA production.
Overall, the data are consistent with the hypothesis that PUAs in
sinking particles affect an increase in remineralization of sinking
particles, which results in a concomitant decrease in the efficiency of
POC export from surface waters. This could in turn lead to retention
of phosphorus and other nutrients in shallower waters, potentially
fueling increased primary productivity on interannual timescales
(Fig. 5). Although PUAs are only a very small component of the
organic carbon in sinking particles (Table S2), their bioactivity exerts
a disproportionate influence on the fate of this carbon in the mesopelagic zone. Our results support a broad-reaching “bioactivity
hypothesis,” which states that the bioactivity of the organic matter
itself, through its ability to stimulate or inhibit particle-associated
bacteria, affects POC export in much the same way that mineral
protection and ballasting affect the efficiency of POC export (47–49).
Testing this hypothesis will involve spatially comprehensive fieldbased research focused on numerous molecular targets, efforts that
must ultimately go far beyond our current study.
ACKNOWLEDGMENTS. We thank H. Fredricks, K. Popendorf, J. Fulton,
L. Sofen, J. Collins, S. Shah, E. Peacock, C. Johnson, C. Loureiro, and A. F. Carvalho
for assistance with sample collection and analysis. We are grateful to J. Ossolinksi
for net trap operations and laboratory analyses. R. Sachdeva, D. Needham, and
J. Fuhrman granted us access to their extensive ARISA ITS fragment length
database. The crews of the R/V Knorr and Atlantic Explorer made invaluable
contributions to our research cruises. This research is funded in part by
the Gordon and Betty Moore Foundation through Grant GBMF3301 (to
B.A.S.V.M., K.D.B., M. Johnson, T. Mincer, and A. Vardi). Additional support
came from National Science Foundation Grants OCE-1031143 (to B.A.S.V.M.)
and OCE-1061883 (to K.D.B. and B.A.S.V.M.) and Office of Naval Research
Grant N000140910091 (to B.A.S.V.M.).
PNAS | May 12, 2015 | vol. 112 | no. 19 | 5913
Materials and Methods
A more detailed explanation of our methods can be found in SI Materials and
Methods. Six iterations of the experiment were conducted overall: SS1, SS2,
SS3, TWNA 1, TWNA2, and SANA1 (Fig. S1 and Table S1). Sinking particles
were collected using unpoisoned surface-tethered net traps deployed at
150 m for 24 h. Before setting up the incubations, the trap material was
diluted by 2-fold (SS2, SS3, TWNA 1, TWNA 2, and SANA 1) or 15-fold (SS1)
with 0.2-μm filtered seawater such that the microbial communities within
the incubations were dominated by particle-associated microbes (∼99%).
Experiments were conducted in triplicate at each station by incubating the
diluted trap material with amendments of varying concentrations of PUAs
(0, 1, 10, and 100 μM) in the dark for 24 h at in situ temperature.
The incubations were conducted in biological oxygen demand bottles
equipped with oxygen optode minisensors (PreSens), which allowed us to monitor the drawdown of O2 during the incubation period and in turn calculate the
respiration rate (50). At the end of the incubations, enzymatic activity of APase,
lipase, α-glucosidase, and aminopeptidase were determined for each triplicate by
measuring the hydrolysis product of commonly used fluorogenic substrates (39).
Bacterial cell abundance was determined using flow cytometry (51). Bacterial
production rates within triplicate treatments were determined by tracing the
uptake of tritiated leucine using the standard microcentrifuge method (52).
A 20-mL sample of each triplicate was taken for PUA derivatization and
extraction at sea. For experiments SS3, TWNA1, TWNA2, and SANA1, t =
initial samples were also extracted. PUAs were quantified using HPLC coupled with UV-visible spectroscopy. Atmospheric pressure chemical ionization
MSn was used to verify that the peaks detected by UV-visible spectroscopy
were indeed the molecules of interest (Fig. S5) (53).
Samples for bacterial community structure analysis were taken at TWNA1
and SS3 by filtering 50 mL of each incubation onto a 0.2-μm-pore size 25-mm
Durapore filter, which were then frozen at –80 °C. Back in the laboratory,
the DNA samples were extracted (54), amplified with 6-FAM–labeled primers
for the 16S–23S intergenic spacer (ITS) region, and community structure was
parameterized using ARISA (55).
Wilcoxon rank-sum tests and Spearman rank-correlation tests were used to
evaluate the statistical significance (P < 0.05) of experimental data, and were
performed with STATISTICA software.
particle specializers (Firmicutes and Bacteriodetes) (Fig. 4 and
Fig. S4).
By using Spearman’s rank correlations to assess the effects of
PUAs on individual bacterial clades, we were able to identify one
PUA-sensitive clade and one PUA-tolerant clade. The negative
correlation between PUA addition and the relative abundance of
γ-Proteobacteria (Table S3), suggests that this group is PUA-sensitive. This is consistent with a report that γ-Proteobacteria became
dominant in the surface community in the North Sea only after a
bloom of potentially PUA-rich diatoms subsided (40). By contrast,
the relative abundance of Actinobacteria was positively correlated
with PUA addition, and this clade was previously shown to be
strongly correlated with diatom pigments in the Sargasso Sea (41).
We found no other correlations between PUA amendments and
any of the other clades we were able to resolve with ARISA. Ribalet
et al. (14) noted that bacterial isolates within the same genera had
the potential to respond differently to PUAs, which suggests that
PUA sensitivity is not strictly defined by taxonomic position.
In contrast to our 100 μM results, recent work by Paul et al.
(42) concluded that PUAs do not impact the structure of freeliving bacterial communities. However, these investigators examined the effects of individual PUAs at nanomolar concentrations,
and it is possible that nanomolar concentrations simply do not
affect community structure, whereas micromolar concentrations
do. In addition, it should be noted that because particle-associated
and free-living bacterial communities are distinct (20, 43), their
corresponding stimulatory and inhibitory concentration ranges are
also likely to be different (14, 44). Alternatively, multiple chain
lengths of PUAs together are often more potent than the same
chain lengths separately, as has been observed for the metabolic
activity of coastal free-living bacteria (16). Diatoms often release
multiple chain lengths of PUAs at once (45), and the accumulation
of heptadienal, octadienal, and decadienal in many of our noamendment controls suggest that this is the case on sinking particles across the North Atlantic (Fig. 3).
The mechanisms by which PUAs stimulate or inhibit marine
bacteria are not known. It has been proposed that PUAs stimulate
bacteria by acting as growth cofactors (14). It is also possible that
PUAs released from diatoms in sinking particles function as cues
for the presence of a larger pool of labile organic matter, which
bacteria respond to with enhanced enzymatic and catabolic activity. We speculate that PUAs might be bona fide signals that
diatoms secrete upon cell death to stimulate remineralization by
bacteria, and thereby increase the likelihood that nutrients such as
P are retained in surface water for the rest of the population to use
(Fig. 5). On the other hand, PUAs are known Michael acceptors,
which are highly reactive toward –NH2 and/or –SH groups, with
the potential to cause unspecific damage inside bacterial cells (46);
Michael reactions could explain PUAs’ inhibitory effects at high
concentrations. Clearly, additional work is necessary to understand the cellular and molecular bases for the impact of PUAs
on marine bacteria in sinking particles.
1. Smith DC, Steward GF, Long RA, Azam F (1995) Bacterial mediation of carbon fluxes
during a diatom bloom in a mesocosm. Deep Sea Res Part II Top Stud Oceanogr 42(1):
2. Azam F, Smith DC, Steward GF, Hagström A (1994) Bacteria-organic matter coupling
and its significance for oceanic carbon cycling. Microb Ecol 28(2):167–179.
3. Wakeham SG, Hedges JI, Lee C, Peterson ML, Hernes PJ (1997) Compositions and
transport of lipid biomarkers through the water column and surficial sediments of the
equatorial Pacific Ocean. Deep Sea Res Part II Top Stud Oceanogr 44(9):2131–2162.
4. Van Mooy BAS, Keil RG, Devol AH (2002) Impact of suboxia on sinking particulate
organic carbon: Enhanced carbon flux and preferential degradation of amino acids
via denitrification. Geochim Cosmochim Acta 66(3):457–465.
5. Goutx M, et al. (2007) Composition and degradation of marine particles with different settling velocities in the northwestern Mediterranean Sea. Limnol Oceanogr
6. Ban S, et al. (1997) The paradox of diatom-copepod interactions. Mar Ecol Prog Ser
7. Ianora A, Poulet SA (1993) Egg viability in the copepod Temora stylifera. Limnol
Oceanogr 38(8):1615–1626.
8. Poulet SA, Laabir M, Ianora A, Miralto A (1995) Reproductive response of Calanus
helgolandicus. I. Abnormal embryonic and naupliar development. Mar Ecol Prog Ser
9. Miralto A, et al. (1999) The insidious effect of diatoms on copepod reproduction.
Nature 402(6758):173–176.
10. Schilmiller AL, Howe GA (2005) Systemic signaling in the wound response. Curr Opin
Plant Biol 8(4):369–377.
11. Wichard T, et al. (2007) Lipid and fatty acid composition of diatoms revisited: Rapid
wound-activated change of food quality parameters influences herbivorous copepod
reproductive success. ChemBioChem 8(10):1146–1153.
12. Casotti R, et al. (2005) Growth inhibition and toxicity of the algal aldehyde 2-trans-2cis decadienal on Thalassiosira weissflogii (Bacillariophyceae). J Phycol 41(1):7–20.
13. Ribalet F, Berges JA, Ianora A, Casotti R (2007) Growth inhibition of cultured marine
phytoplankton by toxic algal-derived polyunsaturated aldehydes. Aquat Toxicol
14. Ribalet F, Intertaglia L, Lebaron P, Casotti R (2008) Differential effect of three polyunsaturated aldehydes on marine bacterial isolates. Aquat Toxicol 86(2):249–255.
15. Ianora A, Miralto A (2010) Toxigenic effects of diatoms on grazers, phytoplankton
and other microbes: A review. Ecotoxicology 19(3):493–511.
16. Balestra C, Alonso-Saez L, Gasol JM, Casotti R (2011) Group-specific effects on coastal
bacterioplankton of polyunsaturated aldehydes produced by diatoms. Aquat Microb
Ecol 63(2):123–131.
17. Vidoudez C, Casotti R, Bastianini M, Pohnert G (2011) Quantification of dissolved and
particulate polyunsaturated aldehydes in the Adriatic sea. Mar Drugs 9(4):500–513.
18. Bartual A, et al. (2014) Polyunsaturated aldehydes from large phytoplankton of the
Atlantic Ocean surface (42°N to 33°S). Mar Drugs 12(2):682–699.
19. Brzezinski MA, Alldredge AL, O’Bryan LM (1997) Silica cycling within marine snow.
Limnol Oceanogr 42(8):1706–1713.
20. DeLong EF, Franks DG, Alldredge AL (1993) Phylogenetic diversity of aggregate-attached vs. free-living marine bacterial assemblages. Limnol Oceanogr 38(5):924–934.
21. Crespo BG, Pommier T, Fernández-Gómez B, Pedrós-Alió C (2013) Taxonomic composition of the particle-attached and free-living bacterial assemblages in the Northwest
Mediterranean Sea analyzed by pyrosequencing of the 16S rRNA. Microbiologyopen
22. Alkire MB, et al. (2012) Estimates of net community production and export using
high-resolution, Lagrangian measurements of O2, NO3−, and POC through the evolution of a spring diatom bloom in the North Atlantic. Deep Sea Res Part I Oceanogr
Res Pap 64:157–174.
23. Martin PLR, Perry MJ, Sanders R (2011) Export and mesopelagic particle flux during a
North Atlantic spring diatom bloom. Deep Sea Res Part I Oceanogr Res Pap 58:
24. Kwon EY, Primeau F, Sarmiento JL (2009) The impact of remineralization depth on
the air-sea carbon balance. Nat Geosci 2(9):630–635.
25. Smith DCSM, Alldredge AL, Azam F (1992) Intense hydrolytic enzyme activity on
marine aggregates and implications for rapid particle dissolution. Nature 359:10.
26. Cochran JK, Buesseler KO, Bacon MP, Livingston HD (1993) Thorium isotopes as indicators of particle dynamics in the upper ocean: Results from the JGOFS North Atlantic Bloom Experiment. Deep Sea Res Part I Oceanogr Res Pap 40(8):1569–1595.
27. Michaels AF, Knap AH (1996) Overview of the US JGOFS Bermuda Atlantic Time-Series
Study and the Hydrostation S program. Deep Sea Res Part II Top Stud Oceanogr 43(2):
28. Suzumura M, Ingall ED (2004) Distribution and dynamics of various forms of phosphorus in seawater: Insights from field observations in the Pacific Ocean and a laboratory experiment. Deep Sea Res Part I Oceanogr Res Pap 51(8):1113–1130.
5914 |
29. Ammerman JW, Hood RR, Case DA, Cotner JB (2003) Phosphorus deficiency in the
Atlantic: An emerging paradigm in oceanography. Eos Trans AGU 84(18):165–170.
30. Teng Y-C, Primeau FW, Moore JK, Lomas MW, Martiny AC (2014) Global-scale variations of the ratios of carbon to phosphorus in exported marine organic matter. Nat
Geosci 7(12):895–898.
31. DeVries T, Deutsch C (2014) Large-scale variations in the stoichiometry of marine
organic matter respiration. Nat Geosci 7(12):890–894.
32. Ribalet F, et al. (2007) Age and nutrient limitation enhance polyunsaturated aldehyde
production in marine diatoms. Phytochemistry 68(15):2059–2067.
33. Ribalet F, et al. (2009) High plasticity in the production of diatom-derived polyunsaturated aldehydes under nutrient limitation: Physiological and ecological implications. Protist 160(3):444–451.
34. Bidle KD, Azam F (1999) Accelerated dissolution of diatom silica by marine bacterial
assemblages. Nature 397(6719):508–512.
35. Bidle KD, Brsesinski MA, Long RA, Jones JL, Azam F (2003) Diminished efficiency in the
oceanic silica pump caused by bacteria-mediatedsilica dissolution. Limnol Oceanogr
36. Ragueneau O, Schultes S, Bidle K, Claquin P, Moriceau B (2006) Si and C interactions in
the world ocean: Importance of ecological processes and implications for the role of
diatoms in the biological pump. Global Biogeochem Cycles 20(4):GB4S02.
37. Leblanc K, et al. (2009) Distribution of calcifying and silicifying phytoplankton in relation to environmental and biogeochemical parameters during the late stages of the
2005 North East Atlantic Spring Bloom. Biogeosciences 6(10):2155–2179.
38. Bidle KD, Azam F (2001) Bacterial control of silicon regeneration from diatom detritus: Significance of bacterial ectohydrolases and species identity. Limnol Oceanogr
39. Hoppe HG (1993) Use of fluorogenic model substrates for extracellular enzyme activity (EEA) measurement of bacteria. Handbook of Methods in Aquatic Microbial
Ecology, eds Kemp PF, Sherr BF, Sherr EB, Cole JJ (Lewis Publishers, Boca Raton, FL), pp
40. Teeling H, et al. (2012) Substrate-controlled succession of marine bacterioplankton
populations induced by a phytoplankton bloom. Science 336(6081):608–611.
41. Nelson CE, Carlson CA, Ewart CS, Halewood ER (2014) Community differentiation and
population enrichment of Sargasso Sea bacterioplankton in the euphotic zone of a
mesoscale mode-water eddy. Environ Microbiol 16(3):871–887.
42. Paul C, et al. (2012) Diatom derived polyunsaturated aldehydes do not structure the
planktonic microbial community in a mesocosm study. Mar Drugs 10(4):775–792.
43. Bidle KD, Fletcher M (1995) Comparison of free-living and particle-associated bacterial communities in the chesapeake bay by stable low-molecular-weight RNA analysis.
Appl Environ Microbiol 61(3):944–952.
44. Amin SA, Parker MS, Armbrust EV (2012) Interactions between diatoms and bacteria.
Microbiol Mol Biol Rev 76(3):667–684.
45. Fontana A, et al. (2007) Chemistry of oxylipin pathways in marine diatoms. Pure Appl
Chem 79(4):481–490.
46. Adolph S, et al. (2004) Cytotoxicity of diatom-derived oxylipins in organisms belonging to different phyla. J Exp Biol 207(Pt 17):2935–2946.
47. Armstrong RA, Lee C, Hedges JI, Honjo S, Wakeham SG (2001) A new, mechanistic
model for organic carbon fluxes in the ocean based on the quantitative association of
POC with ballast minerals. Deep Sea Res Part II Top Stud Oceanogr 49(1):219–236.
48. Francois R, Honjo S, Krishfield R, Manganini S (2002) Factors controlling the flux of
organic carbon to the bathypelagic zone of the ocean. Global Biogeochem Cycles
49. Klaas C, Archer DE (2002) Association of sinking organic matter with various types of
mineral ballast in the deep sea: Implications for the rain ratio. Global Biogeochem
Cycles 16(4):1116.
50. Edwards BR, et al. (2011) Rapid microbial respiration of oil from the Deepwater Horizon spill in offshore surface waters of the Gulf of Mexico. Environ Res Lett 6(3):
51. Tripp HJ (2008) Counting marine microbes with Guava Easy-Cyte 96 well plate reading
flow cytometer. Nat Protoc Exchange. Available at
52. Smith DC, Azam F (1992) A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3H-leucine. Mar Microb Food Webs 6(2):
53. Kölliker S, Oehme M, Dye C (1998) Structure elucidation of 2,4-dinitrophenylhydrazone derivatives of carbonyl compounds in ambient air by HPLC/MS and multiple
MS/MS using atmospheric chemical ionization in the negative ion mode. Anal Chem
54. Santoro AE, Casciotti KL, Francis CA (2010) Activity, abundance and diversity of nitrifying archaea and bacteria in the central California Current. Environ Microbiol
55. Needham DM, et al. (2013) Short-term observations of marine bacterial and viral
communities: Patterns, connections and resilience. ISME J 7(7):1274–1285.
Edwards et al.

Dose-dependent regulation of microbial activity on sinking particles by polyunsaturated aldehydes: